Is P equal to NP? Frank Vega To cite this version: Frank Vega. Is P equal to NP?. 2016. hal-01270398 HAL Id: hal-01270398 https://hal.archives-ouvertes.fr/hal-01270398 Preprint submitted on 7 Feb 2016 HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Public Domain Is P equal to NP? Frank Vega Abstract P versus NP is one of the most important and unsolved problems in computer science. This consists in knowing the answer of the following question: Is P equal to NP? This incognita was first mentioned in a letter written by Kurt Godel¨ to John von Neumann in 1956. However, the precise statement of the P versus NP problem was introduced in 1971 by Stephen Cook in a seminal paper. Under the assumption of P = NP, we show that P = EXP is also hold. Since P is not equal to EXP, we prove that P is not equal to NP by the Reductio ad absurdum rule. Keywords: P, NP, EXP, NEXP, coNP 2000 MSC: 68-XX, 68Qxx, 68Q15 1. Introduction P versus NP is a major unsolved problem in computer science. This problem was introduced in 1971 by Stephen Cook [1]. It is considered by many to be the most important open problem in the field [2]. It is one of the seven Millennium Prize Problems selected by the Clay Mathematics Institute to carry a US$1,000,000 prize for the first correct solution [2]. In 1936, Turing developed his theoretical computational model [3]. The deterministic and nondeterministic Turing machine have become in some of the most important definitions related to this theoretical model for computation. A deterministic Turing machine has only one next action for each step defined in its program or transition function [4]. A nondeterministic Turing machine could contain more than one action defined for each step of its program, where this one is no longer a function, but a relation [4]. Another huge advance in the last century was the definition of a complexity class. A language over an alphabet is any set of strings made up of symbols from that alphabet [5]. A complexity class is a set of problems, which are represented as a language, grouped by measures such as the running time, memory, etc [5]. In computational complexity theory, the class P contains those languages that can be de- cided in polynomial-time by a deterministic Turing machine [6]. The class NP consists in those languages that can be decided in polynomial-time by a nondeterministic Turing machine [6]. The biggest open question in theoretical computer science concerns the relationship between these two classes: Is P equal to NP? Email address: [email protected] (Frank Vega) Preprint submitted to Theoretical Computer Science February 3, 2016 In a 2002 poll of 100 researchers, 61 believed the answer to be no, 9 believed the answer is yes, and 22 were unsure; 8 believed the question may be independent of the currently accepted axioms and so impossible to prove or disprove [7]. Our principal argument is based in a technique that has been used throughout history in both formal mathematical and philosophical reasoning, as well as informal debate: The Reductio ad absurdum [5]. Reductio ad absurdum is a common form of argument which seeks to demonstrate that a statement is true by showing that a false, untenable, or absurd result follows from its denial, or in turn to demonstrate that a statement is false by showing that a false, untenable, or absurd result follows from its acceptance [5]. On the other hand, we have the class EXP contains those languages that can be decided in exponential-time by a deterministic Turing machine [6]. The class NEXP is the set of all languages that can be decided in exponential-time by a nondeterministic Turing machine [6]. EXP and NEXP are nothing else but P and NP on exponentially more succinct input [4]. It is known the succinct version of the problem HAMILTON PATH, that is called SUCCINCT HAMILTON PATH, is in NEXP–complete [4]. We shall prove if we assume that P = NP, then the language SUCCINCT HAMILTON PATH would be in P too. However, this would imply P = EXP [4]. But, this is a false result [4]. In this way, we shall claim that P , NP as a consequence of applying the Reductio ad absurdum rule. 2. Results A graph G is a pair (V; E), where V is a finite set and E is a binary relation on V [5]. The set V is called the vertex set of G, and its elements are called vertices or nodes [5]. The set E is called the edge set of G, and its elements are called edges [5]. If (u; v) is an edge in a graph G = (V; E), we say that vertex v is adjacent to vertex u [5]. A path of length k from a vertex u 0 to a vertex u in a graph G = (V; E) is a sequence of vertices hv0; v1; v2;:::; vki such that u = v0, 0 u = vk, and (vi−1; vi) 2 E for i = 1; 2;:::; k [5]. One of the most basic problems on graphs is this: Given a graph, is there a path that visits each node exactly once? We call this problem as HAMILTON PATH [4]. HAMILTON PATH is in NP–complete [4]. A succinct representation of a graph with n nodes, where n = 2b is a power of two, is a Boolean circuit C with 2 × b input gates [4]. The graph represented by C, denoted GC, is defined as follows: The nodes of GC are f1; 2;:::; ng. And (i; j) is an edge of GC if and only if C accepts the binary representations of the b-bits integers i, j as inputs [4]. In addition, we represent the number n with the b-bits integer 0. The problem SUCCINCT HAMILTON PATH is now this: Given the succinct representation C of a graph GC with n nodes, does GC have a Hamilton path? The problem SUCCINCT HAMILTON PATH is in NEXP–complete [4]. Theorem 2.1. If P = NP, then SUCCINCT HAMILTON PATH would be in P. b Proof. Let’s take an arbitrary succinct representation C of a graph GC with n nodes, where n = 2 is a power of two and C will be a Boolean circuit of 2 × b input gates. The circuit C computes 2×b a Boolean function fC : ftrue; f alseg ! ftrue; f alseg [4]. Now, if C is a “yes” instance of SUCCINCT HAMILTON PATH, then there will be a linear order Q on the nodes of GC, that is, a binary relationship isomorphic to < on the nodes of GC, such that consecutive nodes are connected in GC [4]. This linear order Q must require several things: 1. All distinct nodes of GC are comparable by Q, 2 2. next, Q must be transitive but not reflexive, 3. and finally, any two consecutive nodes in Q must be adjacent in GC. Any binary relationship Q that has these properties must be a linear order, any two consecu- tive elements of which are adjacent in GC, that is, it must be a Hamilton path [4]. Let R be a binary relation on strings. R is called polynomially decidable if there is a deter- ministic Turing machine deciding the language fx; y :(x; y) 2 Rg in polynomial-time [4]. We say that R is polynomially balanced if (x; y) 2 R implies jyj < jxjk for some k ≥ 1 [4]. The linear order Q can be represented as a graph GQ. In this way, the succinct representation CQ of the graph GQ will represent the linear order Q too. We can define a polynomially balanced relation RQ, where for all succinct representation C of a graph: There is another Boolean circuit CQ that will represent a linear order Q on the nodes of GC such that (C; CQ) 2 RQ if and only if C 2 SUCCINCT HAMILTON PATH [4]. Indeed, the graphs GC and GQ will comply with 3 jGQj < jGCj when (C; CQ) 2 RQ, since both graphs would have the same number of nodes and GC would contain a Hamilton path. Certainly, if the graph GC of n nodes contains a Hamilton path, then it would have at least (n − 1) edges. But, if GC is a pair (VGC ; EGC ), GQ is (VGQ ; EGQ ) and (C; CQ) 2 RQ, where VGC and VGQ are vertex sets and EGQ and EGC are edge sets, then 3 jVGC j = jVGQ j and jEGQ j < jEGC j when jEGC j > 1, because the maximum number of edges in a graph of n nodes is lesser than n × (n − 1) [5]. Consequently, we obtain the same property for their succinct representations, that is, CQ should be polynomially bounded by C. Indeed, for a sufficiently large n, the Boolean circuits C and CQ will be exponentially more succinct than GC and GQ respectively [4]. Hence, if the graph GQ is polynomially bounded by the graph GC when (C; CQ) 2 RQ, then log2 jGQj will be polynomially bounded by log2 jGCj.